Neuroimaging transcriptomic analyses of Parkinson's disease highlight molecular, cellular, and neurobiological mechanisms.

Parkinson's disease is a debilitating disorder with a profound symptomatic burden. While genomic studies of PD have made advances in recent years, there remains a gap in our understanding of the relationship between PD genes and the neurologic changes involved in disease progression. In this work, we aim to characterize the mechanisms by which germline susceptibility variants affect gene expression and the downstream effects on neurobiology. We perform a transcriptome-wide association study of PD (N = 482,730) and identify 74 transcript-level associations with the disease. These findings derive from gene expression models trained in bulk tissue samples from 13 different brain regions, and 6 tissues with clinical relevance to the brain, including whole blood. We explore the resulting associations using eQTL colocalization, the chromatin-state discovery approach CoRE-BED, and our NeuroimaGene atlas of neuroimaging transcriptomics. In this work, we implicate enhancer disruption for STX1B and HIP1R in neurons and oligodendrocytes respectively in PD pathophysiology and show that reduced expression of PD risk genes broadly is predictive of cortical thinning and dysregulation of somatomotor circuitry.